Background and Purpose The purposes of this study were to examine overall and domain-specific quality of life in long-term stroke survivors and to identify variables that predict quality of life after stroke.
Methods The study used a cross-sectional, descriptive correlational design. Subjects were 86 stroke survivors who were interviewed 1 to 3 years after stroke. Quality of life was measured with the use of an instrument that assesses satisfaction and importance for four domains (health and functioning, socioeconomic, family, and psychological-spiritual). Independent variables were age, social class, aphasia, functional status, motor impairment, depression, comorbidity, and perceived social support. Multiple regression analysis was used to predict quality of life.
Results Thirty percent of subjects scored in the depressed range. The mean overall quality of life score was relatively high and was comparable to that of a normative population. Quality of life was highest for the family domain and lowest for health and functioning. Depression, perceived social support, and functional status predicted quality of life (adjusted R2=.38) and contributed to prediction of domain quality of life. Social support and three additional variables, social class, age, and cardiovascular disease, predicted socioeconomic quality of life.
Conclusions The identification of depression, social support, and functional status as predictors of quality of life suggests the need to assist stroke survivors in coping and in maintaining and strengthening their support systems. (Stroke. 1996;27:1467-1472.)
There is a growing consensus that QOL is an important healthcare outcome.1 2 3 Despite the potential of stroke to affect multiple domains of life and related outcomes, few American studies have addressed QOL after stroke. The studies that have been reported used either a single item or assessment of three to four items considered to represent QOL or life satisfaction.4 5 6 7 However, a multidimensional approach is needed to provide direction in improving QOL.8 European investigators using such an approach reported diminished global, leisure, sexual, and work-life satisfaction.9 10
Gill and Feinstein1 proposed that “the need to incorporate patients' values and preferences is what distinguishes quality of life from all other measures of health.” QOL studies have rarely included ratings of item importance, and no such study of stroke survivors could be found.
Variables reported to be associated consistently with lower QOL or life satisfaction in stroke survivors are depression,9 11 12 lower functional status,4 9 11 12 and greater severity of paralysis.9 10 In one of these studies,4 an indirect relationship was reported between admission function and life satisfaction. Inconsistent findings have been reported for age.4 9 10 11 13 Other variables such as social support, comorbidity, and aphasia have been examined rarely. A single report on aphasia indicated no significant relationship to QOL.9 Reports have indicated that social support or social contacts in stroke survivors were positively and significantly related to domain and global QOL.4 10 A greater number of comorbid conditions has been reported to be related to QOL.13 Although SES was related to QOL in a study of older persons,14 research on stroke survivors has shown no relationship11 or a weak, indirect association.7
Interpretation of QOL findings is complicated by measurement issues such as simultaneous assessment of prestroke and poststroke QOL,9 10 11 use of nonstandardized measures of depression,9 11 and lack of control for duration of stroke in samples with wide variability in time after onset.15 The present study addressed these issues and used a multidimensional measure that included assessment of item importance.16
The purposes of this study were to examine global and domain-specific QOL in individuals with long-term stroke and to examine the relationship between selected variables and QOL. Domain-specific findings and knowledge of QOL predictors can provide support for interventions to improve QOL.
Subjects and Methods
A cross-sectional, descriptive design was used. Subjects were identified from records of consecutive discharges of 698 stroke survivors during a 2-year period from site A (a freestanding rehabilitation hospital) and during a 16-month period from site B (a community teaching hospital). Both sites provided acute multidisciplinary rehabilitation, with patients admitted to one of two 38-bed units at site A or to a 26-bed unit at site B. Inclusion criteria were the following: (1) the patient had been discharged for 1 to 3 years from a rehabilitation program for first completed stroke (or had no residual impairment from a previous stroke, if a recurrent stroke), (2) absence of other neuromusculoskeletal condition that interfered with function, (3) absence of previous psychiatric history, (4) cognitive and language function adequate to participate in study procedures, and (5) residing in a noninstitutional setting. Diagnosis of stroke was confirmed by clinical and/or radiological findings. From a pool of 121 persons who met the inclusion criteria and could be located, 86 (71%) agreed to participate. Marital status, age, sex, and duration of stroke were not significantly different (P>.05) between participants and nonparticipants.
Cognitive status and language were screened to determine ability to respond reliably to interview questions; 6 subjects who consented but did not meet screening criteria were not interviewed. Characteristics of 86 subjects from the two sites were equivalent except for age, which was greater in subjects from the community hospital (72.5 versus 65.5 years old), t=2.61 (df=84), P=.02.
Cognitive status was measured by use of the MMSE, which possesses high criterion validity17 and high sensitivity and specificity.18 For the present study, an error of one in identifying day or date was acceptable. The recommended cutoff of ≥24 was used to select subjects. Aphasia was measured by use of the Language Screening Test,19 which measures auditory comprehension and oral expression. On the basis of the results of the Language Screening Test, aphasia severity was rated by use of the Severity Impairment Scale.19 Potential subjects with mild to moderate aphasia met the study criteria. Aphasia, a predictor variable, was coded as absent (0) or present (1) for analysis because severity data were skewed.
QOL was defined as satisfaction with aspects of life that are important to the individual.20 Therefore, QOL was measured by use of the Ferrans & Powers Quality of Life Index-Stroke Version (QLI),16 which is a two-part scale that rates 38 items for satisfaction (part 1) and importance (part 2). Likert-scaled responses range from 1 (very dissatisfied/very unimportant) to 6 (very satisfied/very important). Factor analysis identified four subscales: health and functioning, socioeconomic, psychological-spiritual, and family.20 The original QLI was modified by the addition of communication, self-care, and mobility items. The QLI possesses strong internal consistency (α=.90 to .93) and concurrent validity.16 The possible range of scores for subscales and overall QLI is 0 to 30; lower scores indicate lower QOL. The scoring procedure is provided in the Appendix.
Data were collected on sex, marital status, race, education, duration of stroke, and location of brain lesion for descriptive purposes. In addition to aphasia, independent variables included age, SES, comorbid conditions, functional status, motor impairment, depression, and perceived social support. SES was measured by use of the Hollingshead Four-Factor Index,21 which estimates social class. Scores range from 8 (menial work) to 66 (professional and major business work).
Data on comorbid conditions were retrieved from the rehabilitation service medical records. In addition, subjects were asked to identify conditions diagnosed since the last medical record entry. Self-report was used to supplement comorbidity data because of the inability to access records maintained by the subjects' current healthcare providers. Strong agreement has been reported between self-reported chronic diseases and physician records.22 Absence of a comorbid condition was coded as 0, and presence was coded as 1.
Functional status was measured by use of a 15-item version of the FIM and a four-point scoring system.23 Three social cognition items were not included because of the difficulty in obtaining a reliable measure during one interaction. Responses for each item ranged from “complete dependence” (1) to “complete independence” (no helper or assistive device needed; reasonable time required to complete tasks) (4). The possible range of scores is 15 to 60. Intraclass correlations have been reported to range from .86 to .88.23
Motor impairment was estimated by the investigators and dichotomized. A rating of 1 indicated functional use of extremities. The presence of one or more nonfunctional extremities was coded as 2.
Depression was measured by use of the 20-item CES-D,24 which measures depressive symptoms on a four-point scale. The range of scores is 0 to 60, with higher scores indicating greater severity of symptoms; a score ≥16 is indicative of depression. Coefficient alphas of .85 to .9024 and strong correlations with other measures of depression25 have been reported.
Perceived social support was measured by use of the SSE (M. Powers, PhD, RN, and J. Miller, PhD, RN, unpublished data, 1986), which rates the satisfaction and importance of support. This nine-item scale assesses tangible, informational, and emotional support and possesses a coefficient alpha of .83.26 Scores are computed by use of the QLI scoring program16 and range from 0 to 30, with higher scores indicating greater perceived support.
The study was approved by the Institutional Review Boards at the two sites. Informed consent was obtained at the time of data collection. All but five interviews were conducted in the subjects' homes. The author and a masters-prepared nurse collected data. Instruments were read to participants, and responses were recorded by the investigators. Interrater agreement was 93% for the FIM and 100% for the MMSE and motor impairment, using data collected on five subjects.
Student's t test, the χ2 statistic, and correlation coefficients were used to compare subjects from the two sites and to examine relations between study variables. Stepwise multiple regression was computed to predict QOL. Dichotomous variables were included in the regression analysis if at least an 80%/20% dichotomy was found.27 A two-tailed probability value <.05 was the criterion for statistical significance.
The majority of the sample was female (55%), married (62%), white (78%), and had graduated from high school (69%). The median age was 69.0 years (63.3±13.2 [mean±SD]; range, 23 to 88 years). Illness characteristics (hemispheric side of lesion, duration of stroke, language impairment, motor impairment, and comorbidity) are shown in Table 1⇓. Four comorbid conditions (cardiovascular disease, hypertension, diabetes, and arthritis) and aphasia were found in ≥20% of the subjects.
Descriptive statistics and reliability coefficients for SES, FIM, CES-D, SSE, and overall and domain subscale scores on the QLI are presented in Table 2⇓. The mean SES score corresponds to a skilled worker. Depression was common, with 30% scoring in the depressed range on the CES-D. The mean SSE score indicated that subjects perceived their social support to be high.
The mean and median FIM scores reflected modified independence such that an assistive device or extended time was required. Physical assistance was required by 43% of the subjects for one or more self-care or mobility functions. Logarithmic transformation was used to improve the distribution of the skewed FIM scores after the scores were reflexed to produce a positive skew.27 With the transformed score, higher values represented lower function. The new variable, FIMLOG (0.82±0.35), was used in regression analysis.
The first purpose of the present study was to assess overall and subscale QOL. Mean QOL scores were relatively high (Table 2⇑). Mean QLI subscale scores ranged from 21.5 (health and functioning) to 26.9 (family). Table 3⇓ shows the 10 highest- and lowest-weighted QOL item scores. Among items with the highest QOL rankings are several that relate to family and friends. To facilitate comparisons with other studies, findings for the item, “How satisfied are you with your life in general?” were examined. The median (5.0) reflected moderate satisfaction. Nevertheless, 23% of the sample reported dissatisfaction or only slight satisfaction with their lives.
In general, participants rated items as moderately or very important. However, the following items received median ratings of slightly important or unimportant: (1) to have work (asked only of those not working), (2) to have influence on government, (3) to live a long time, and (4) to be able to travel. Individual differences in values were found when several health and functioning items were compared. Whereas 99% of participants rated health care as moderately or very important, only 69% and 75%, respectively, gave similar ratings for sex life and travel.
Similarities and differences were found in variables predicting domain and overall QOL. Table 4⇓ shows results of the stepwise multiple regression procedures. The family domain was not analyzed because of a low reliability coefficient. Three variables explained 38% (adjusted R2) of the variance in QOL. The predictors of overall QOL were depression (β=−.53, P=.0001), perceived social support (β=.33, P=.0005), and FIMLOG (β=−.19, P=.031). One or more of these variables also predicted psychological, health and functioning, and socioeconomic QOL. Three variables (SES, cardiovascular disease, and age), in addition to social support, predicted socioeconomic QOL. The relations were such that greater depression, less social support, lower functional status, presence of cardiovascular disease, younger age, and lower SES were associated with lower overall and/or domain QOL. Five illness variables (arthritis, diabetes, hypertension, motor impairment, and aphasia) were not predictors of QOL.
The relatively high QOL found in this sample suggests that most subjects were coping effectively with stroke-related impairments. Overall QOL was similar to the mean QLI score of 23.0 found in a sample of 339 subjects drawn randomly from a telephone directory. The sample (mean age=48 years) included urban, suburban, and rural communities (personal communication with C.E. Ferrans, PhD, RN, coauthor of the QLI, March 20, 1996). The findings can be compared with a sample of elderly outpatients with chronic illnesses, who perceived their QOL as good.28
Because other studies of stroke survivors compared past with present QOL9 11 or did not provide a relative estimate of QOL,7 comparisons with other stroke QOL findings are limited. However, the finding of relatively high QOL is consistent with reports indicating that 52% to 82% of long-term stroke survivors were satisfied with their lives.5 10 12 Despite the high mean QOL score, a subsample of 23% reported low life satisfaction in response to a single item on satisfaction with their lives. Perceptions of a high level of QOL for family items (Table 3⇑) contributed to the relatively high overall QOL. The mean family domain score was higher than those of other subscales. Findings from the present study and others must be interpreted in light of the selection criteria, which excluded individuals residing in institutional settings and those with severe cognitive and language impairments, who may have lower QOL.
One explanation that was considered for the relatively high QOL was the long-term duration of the stroke. Wright29 suggested that with time, coping strategies (eg, enlarging the scope of values) contribute positively to the meaning of life with a disability. However, Ahlsio et al11 found no significant change in QOL from short-term care through 2 years after stroke. The correlation between duration of stroke and QOL was not significant in the present study (r=−.13), and the direction of the relation contrasted with Wright's suggestion.
Several items that ranked low in QOL may be amenable to interventions. For instance, counseling to promote healthy coping could include discussion of perceptions of control, stress management, and guidance in reappraising perceptions of being useful and ways of enjoying life. Family members play a critical role in promoting such behavior change and must be included in interventions to facilitate healthy coping. Changing aspirations and values after a stroke is a long-term process, and thus, follow-up assessments can be important in improving QOL.
The finding of low quality of sexual life (Table 3⇑) is similar to a report that indicated a significant decrease in sexual satisfaction after stroke.10 In the present study, many subjects elaborated on sexual difficulties when they responded to the QLI satisfaction item on sexual life. For example, an 80-year-old male discussed his distress concerning erectile problems that occurred after the stroke. Like several other subjects with low sexual QOL, he had not discussed this issue with a healthcare professional. Assessment of sexual concerns with appropriate education and counseling should be included in follow-up assessments.
The finding that not having a job is a major source of dissatisfaction for a subsample in the present study is congruent with findings in a sample of subjects <65 years old9 and a national study in which nonworkers with a disability reported lower life satisfaction than their working counterparts.30 Because working is an important aspect of life for many individuals, counseling on finding purpose in other roles may be beneficial for those who cannot return to work.
The second purpose of the present study was to examine the relationship between selected variables and QOL. Depression was the strongest predictor of overall, psychological, and health QOL. This finding supports the results of several studies9 11 12 but contrasts with that of Viitanen et al.10 The latter report indicated a 16% prevalence of depression. The longer duration of stroke in the study by Viitanen and colleagues10 may account for some of the difference. The depression prevalence rate in the present study is similar to other reports for a similar duration of stroke.11 31 32 The high rate of depression supports the need for follow-up programs that include assessment and interventions to treat depression.
Social support emerged as an important predictor and was the only variable that was a predictor of overall QOL and QOL for each domain. Other investigators have found a relationship between less social contact or support and emotional maladaptation33 and lower life satisfaction10 in stroke survivors. Referrals of survivors to support groups and education of family members and the wider community on the importance of social support after disability may help to strengthen support. In addition, many caregivers benefit from counseling on ways of managing their responses to stroke, so they can continue to be supportive over the long term.
Similar to two studies in which function was assessed at the time of QOL assessment,10 11 function explained little (3%) variance in overall QOL. These findings contrast with a report9 in which difficulty in ambulation and activities of daily living were important predictors of QOL. Methodological issues may account for differences. In that study,9 the change in QOL was examined, not relative scores, and ambulation was assessed by use of a single item. In the current study, function did predict health and functioning and overall QOL but was not related to psychological or socioeconomic QOL. The fact that function and other predictors did not explain more of the variance in QOL supports the need to examine other variables. Impairments such as dysarthria and reduced balance and coordination that were present in this sample but were not study variables may contribute to lower QOL. Other variables, such as optimism, cognitive appraisal of the significance of the stroke, coping skills, and family functioning, have been examined rarely and hold the promise of explaining differences in QOL.
Theories on stress, coping, and adaptation to illness34 35 suggest that cognitive appraisal of the significance of illness and the ability to cope with stressful events influence coping responses and outcomes. Future research that explores the relationship between these variables and QOL is needed to address the complexity of QOL. For example, a report indicated that the use of avoidance coping by the stroke survivor and spouse was correlated with lower life satisfaction in a sample of 73 stroke survivors.36
Although older age, higher social class, and absence of cardiovascular disease contributed to higher socioeconomic QOL, they did not predict overall QOL directly. The finding concerning age appears to contrast with reports that showed that global QOL decreased more for older stroke survivors.9 13
Variables that did not predict either overall or subscale QOL were aphasia, motor impairment, diabetes, arthritis, and hypertension. Findings concerning motor impairment contrast with two studies that reported a relationship between paralysis and QOL.9 10 Dichotomous assessment of motor impairment in the current study may not have provided adequate discrimination between subjects to detect a significant relationship. Niemi et al9 reported no relationship between aphasia and QOL. In the current study, the ability to communicate was rated as highly important (ninth most important of 38 items). Given its importance, the contribution of aphasia to QOL warrants further study.
A limitation of this study was the lack of a comparison group of healthy adults. However, it was possible to compare findings based on the same measure to a normed sample. Convenience sampling and exclusion of persons with severe cognitive or language deficits limit the ability to generalize the findings. The cross-sectional design used in this study prevented examination of changes in QOL over time.
Findings on variability in QOL item importance underscore the need to assess what matters to each stroke survivor and to tailor programs on the basis of their responses. A limitation in the stroke QOL literature is that no two studies have used the same measure. Identification of an instrument that is widely accepted by stroke researchers is needed to advance the study of QOL. A longitudinal study that examines QOL from short-term rehabilitation through long-term adaptation is needed to provide data on fluctuations in QOL and its predictors.
Selected Abbreviations and Acronyms
|CES-D||=||Center for Epidemiologic Studies–Depression Scale|
|FIM||=||Functional Independence Measure|
|MMSE||=||Mini-Mental State Examination|
|QLI||=||Ferrans and Powers Quality of Life Index–Stroke Version|
|QOL||=||quality of life|
|SSE||=||Social Support in the Elderly Scale|
Scoring Procedure for the QLI
Overall and subscale QLI scores were computed by use of the established procedure.16 Each satisfaction response was weighted with its corresponding importance rating. Weighting was done by subtracting 3.5 from each satisfaction response to center the satisfaction scale on 0. This procedure made 0 the midpoint. Each recoded satisfaction item score was multiplied by its paired importance score (possible range=−15.0 to 15.0). To calculate overall scores, weighted items were summed and divided by the number of items answered. To eliminate negative values, a constant of 15 was added to calculate the final score. The possible range of overall QLI scores was 0 to 30. Subscale scores were computed by application of the same scoring procedure to items in each subscale.
This research was funded in part by a predoctoral fellowship from the National Center for Nursing Research, No. NRO6171-02. The author expresses gratitude to Dr Marjorie Powers for guidance in conducting this research and Drs Carolyn Carlson, Elliot Roth, and Allen Heinemann for their helpful suggestions on the manuscript.
- Received January 25, 1996.
- Revision received June 11, 1996.
- Accepted June 11, 1996.
- Copyright © 1996 by American Heart Association
DeJong G, Branch LG. Predicting the stroke patient's ability to live independently. Stroke. 1982;13:648-655.
Osberg JS, McGinnis GE, DeJong G, Seward ML. Life satisfaction and quality of life among disabled elderly adults. J Gerontol. 1987;42:228-230.
de Haan R, Aaronson N, Limburg M, Langton Hewer R, van Crevel H. Measuring quality of life in stroke. Stroke. 1993;24:320-327.
Niemi M, Laaksonen R, Kotila M, Waltimo O. Quality of life 4 years after stroke. Stroke. 1988;19:1101-1107.
Ahlsio B, Britton M, Murray V, Theorell T. Disablement and quality of life after stroke. Stroke. 1984;15:886-890.
Astrom M, Asplund K, Astrom T. Psychosocial function and life satisfaction after stroke. Stroke. 1992;23:527-531.
de Haan RJ, Limburg M, Van der Muelen JHP, Jacobs HM, Aaronson NK. Quality of life after stroke: impact of stroke type and lesion location. Stroke. 1995;26:402-408.
George L, Bearon L. Quality of Life in Older Persons. New York, NY: Human Sciences Press; 1980.
Halper A, Mogil S, Full M, Keatings C, Mito M, Otto K. Aphasia Diagnostic Classifications Severity Ratings. Chicago, Ill: Rehabilitation Institute of Chicago, Department of Communicative Disorders; 1982.
Hollingshead AB. Four-Factor Index of Social Status. New Haven, Conn: Yale University Department of Sociology; 1975.
Granger CV, Hamilton BB, Sherwin FS. Guide for Use of the Uniform Data Set for Medical Rehabilitation. Buffalo, NY: Department of Rehabilitation Medicine, Buffalo General Hospital; 1986.
Radloff LS, Locke B. The community mental health assessment survey and the CES-D scale. In: Weissman M, ed. Community Surveys. New York, NY: Prodist Publisher; 1986:177-189.
Shinar D, Gross C, Price TR, Banko M, Bolduc P, Robinson RG. Screening for depression in stroke patients: the reliability and validity of the Center for Epidemiologic Studies–Depression Scale. Stroke. 1986;17:241-245.
Murray-Finis N. The Elderly Patient's Use of Nonurgent Emergency Care. Chicago, Ill: University of Illinois at Chicago; 1988. Thesis.
Tabachnik B, Fidell L. Using Multivariate Statistics. New York, NY: Harper & Row; 1983.
Pearlman RA, Uhlmann RF. Quality of life in chronic disease: perceptions of elderly patients. J Gerontol. 1988;43:M25-M30.
Wright BA. Physical Disability: A Psychological Approach. New York, NY: Harper & Row; 1960.
The ICD Survey of Disabled Americans: Bringing Disabled Americans Into the Mainstream. New York, NY: International Center for the Disabled; 1986.
Wade DT, Legh-Smith J, Hewer RA. Depressed mood after stroke. Br J Psychiatry. 1987;151:200-205.
Lazarus RS, Folkman S. Stress, Appraisal, and Coping. New York, NY: Springer Publishing Co Inc; 1984.
Moos R, Tsu V. The crisis of physical illness: an overview. In: Moos RH, ed. Coping With Physical Illness. New York, NY: Plenum Press; 1977:3-21.
Robinson-Smith G. Coping and life satisfaction after stroke. J Stroke Cerebrovasc Dis. 1993;3:209-215.